| Commit message (Collapse) | Author | Age | Files | Lines |
|\
| |
| | |
SI-8625 fix unreachability analysis for boolean expressions
|
| | |
|
|\ \
| |/
|/| |
Compiler optimizations for Scopes, checkDoubleDefs, Namer
|
| |
| |
| |
| |
| |
| |
| | |
- No need to check the result type, as dependent method types
are now enabled unconditionally.
- This also means we can only need to check methods with two or
more parameter lists.
|
| |
| |
| |
| |
| |
| |
| | |
Either direct defaults, or directly inherited.
The avoids `addDefaultGetters` in most cases, and the expensive
duplication of trees therein.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Where N is the number of members of the enclosing package.
Double definition errors for top level classes/objects are issued
elsewhere, as demonstrated by the enclosed test. So we can
omit the call to `checkNoDoubleDefs` in this content.
We can't omit the call to `addSynthetics` for package
class owners (case- and value-class synthetic companions
are added here), but we can make the process cheaper by
moving the expensive-but-usually-true call to `shouldAdd`.
Here's an example of the improvement.
% rm -rf /tmp/pkg; (for i in {1..50}; do for j in {1..100}; do echo "package pkg { class A_${i}_${j}___(val a: Int, val b: Int) }"; done; done) > sandbox/A1.scala && time scalac-hash v2.11.0 -Ybackend:GenASM -J-Xmx1G -J-XX:MaxPermSize=400M -d /tmp sandbox/A1.scala;
real 0m49.762s
user 1m12.376s
sys 0m2.371s
% rm -rf /tmp/pkg; (for i in {1..50}; do for j in {1..100}; do echo "package pkg { class A_${i}_${j}___(val a: Int, val b: Int) }"; done; done) > sandbox/A1.scala && time qbin/scalac -Ybackend:GenASM -J-Xmx1G -J-XX:MaxPermSize=400M -d /tmp sandbox/A1.scala;
real 0m35.662s
user 0m58.275s
sys 0m2.355s
We've still got another source of pathological performance in
creating nested scopes that I'll fix in the next commit.
|
|/
|
|
| |
The tree to create a `NoManifest` was unpositioned.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit is a squashed version of all commits in PR #3747.
For future upgrades, consult the README and check the commits in
https://github.com/scala/scala/pull/3747/commits
There's one bug in ASM 5.0.2 that breaks scalac:
http://forge.ow2.org/tracker/?func=detail&aid=317200&group_id=23&atid=100023
This bug is fixed in ASM trunk, the patch has been merged into this
commit. A future upgrade of ASM should contain the fix.
|
|
|
|
| |
Otherwise we lose the side effect of a `NegativeArraySizeException`.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It can NPE or trigger static class initilization, we can't elimiate
it without changing semantics.
To make sure we don't thwart closure elimination, I've allowed DCE
to eliminate a non-static LOAD_FIELD of a member of a closure class.
It would be more general to track nullity of the reciever (e.g, `this`
or `new Foo` cannot be null), but that would require more
infrastructure in this phase.
I've added a test for closure inlining based on a a suggestion by
@dragos. This actually passes if we remove the (LOAD_FIELD, DROP)
peephole optimization for `closelim` altogether. But I chose to
adapt that optimization (only allow it for non-static, closure
fields), rather then remove it alogether, in the interests of
treading lightly.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
`{i, l}div` and `{i, l}rem` throw an `ArithmeticException` if the
divisor is 0.
`arraylength` throws a `NullPointerException` on a null reference.
JVM Spec:
> The only integer operations that can throw an exception are the
> integer divide instructions (idiv and ldiv) and the integer
> remainder instructions (irem and lrem), which throw an
> ArithmeticException if the divisor is zero.
> The Java virtual machine's floating-point operators do not throw
> runtime exceptions
> If the arrayref is null, the arraylength instruction throws a
> NullPointerException.
I checked the other primitives in `ICode` to see if anything else
should be considered as live code.
Pure:
// jvm : {i, l, f, d}neg
case class Negation(kind: TypeKind) extends Primitive
// jvm : if{eq, ne, lt, ge, le, gt}, if{null, nonnull}
// if_icmp{eq, ne, lt, ge, le, gt}, if_acmp{eq,ne}
case class Test(op: TestOp, kind: TypeKind, zero: Boolean) extends Primitive
// jvm : lcmp, {f, d}cmp{l, g}
case class Comparison(op: ComparisonOp, kind: TypeKind) extends Primitive
Impure: {i, l}{div, rem}, otherwise pure
// jvm : {i, l, f, d}{add, sub, mul, div, rem}
case class Arithmetic(op: ArithmeticOp, kind: TypeKind) extends Primitive
Pure (overflow is silent, NaN.toInt is defined):
// jvm : {i, l}{and, or, xor}
case class Logical(op: LogicalOp, kind: TypeKind) extends Primitive
// jvm : {i, l}{shl, ushl, shr}
case class Shift(op: ShiftOp, kind: TypeKind) extends Primitive
// jvm : i2{l, f, d}, l2{i, f, d}, f2{i, l, d}, d2{i, l, f}, i2{b, c, s}
case class Conversion(src: TypeKind, dst: TypeKind) extends Primitive
Impure! May NPE!
// jvm : arraylength
case class ArrayLength(kind: TypeKind) extends Primitive
Pure (we know that StringBuilder.{<init>, append, toString} are pure
and `append` is null safe.)
// jvm : It should call the appropiate 'append' method on StringBuffer
case class StringConcat(el: TypeKind) extends Primitive
// jvm: it should create a new StringBuffer
case object StartConcat extends Primitive
// jvm: convert StringBuffer to a String
case object EndConcat extends Primitive
|
|
|
|
| |
This reverts commit ee611cd76c29fedd416162e482c7ab3f15b831ca.
|
|
|
|
| |
This reverts commit 0b432f9cd22b6e9770852e5b331a15f0534a312c.
|
|
|
|
| |
This reverts commit 70b912a87433c9589af33e4f8b33dca39abb66e5.
|
|
|
|
| |
Otherwise we lose the side effect of a `NegativeArraySizeException`.
|
|
|
|
|
| |
It can NPE or trigger static class initilization, we can't elimiate
it without changing semantics.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
`{i, l}div` and `{i, l}rem` throw an `ArithmeticException` if the
divisor is 0.
`arraylength` throws a `NullPointerException` on a null reference.
JVM Spec:
> The only integer operations that can throw an exception are the
> integer divide instructions (idiv and ldiv) and the integer
> remainder instructions (irem and lrem), which throw an
> ArithmeticException if the divisor is zero.
> The Java virtual machine's floating-point operators do not throw
> runtime exceptions
> If the arrayref is null, the arraylength instruction throws a
> NullPointerException.
I checked the other primitives in `ICode` to see if anything else
should be considered as live code.
Pure:
// jvm : {i, l, f, d}neg
case class Negation(kind: TypeKind) extends Primitive
// jvm : if{eq, ne, lt, ge, le, gt}, if{null, nonnull}
// if_icmp{eq, ne, lt, ge, le, gt}, if_acmp{eq,ne}
case class Test(op: TestOp, kind: TypeKind, zero: Boolean) extends Primitive
// jvm : lcmp, {f, d}cmp{l, g}
case class Comparison(op: ComparisonOp, kind: TypeKind) extends Primitive
Impure: {i, l}{div, rem}, otherwise pure
// jvm : {i, l, f, d}{add, sub, mul, div, rem}
case class Arithmetic(op: ArithmeticOp, kind: TypeKind) extends Primitive
Pure (overflow is silent, NaN.toInt is defined):
// jvm : {i, l}{and, or, xor}
case class Logical(op: LogicalOp, kind: TypeKind) extends Primitive
// jvm : {i, l}{shl, ushl, shr}
case class Shift(op: ShiftOp, kind: TypeKind) extends Primitive
// jvm : i2{l, f, d}, l2{i, f, d}, f2{i, l, d}, d2{i, l, f}, i2{b, c, s}
case class Conversion(src: TypeKind, dst: TypeKind) extends Primitive
Impure! May NPE!
// jvm : arraylength
case class ArrayLength(kind: TypeKind) extends Primitive
Pure (we know that StringBuilder.{<init>, append, toString} are pure
and `append` is null safe.)
// jvm : It should call the appropiate 'append' method on StringBuffer
case class StringConcat(el: TypeKind) extends Primitive
// jvm: it should create a new StringBuffer
case object StartConcat extends Primitive
// jvm: convert StringBuffer to a String
case object EndConcat extends Primitive
|
|\
| |
| | |
SI-8574 Copy @SerialVersionUID, etc, to specialized subclasses
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The test case demonstrates that this is important for serialization
and for strictfp. (Although the latter is still pretty broken,
see SI-7954.)
Now that the synthetic subclass of `Tuple2[Int, Int]` also has the
`@deprecatedInheritance` annotation, I had to change the spot that
issues this warning to be silent after the typer phase. Otherwise,
we get two warnings in `run/t3888.scala`. This also remedies double
warnings that were incurred in `neg/t6162-inheritance`.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
I removed the `-bcode` test since we have a build that passes
`-Ybackend:GenBCode` to all tests.
Short intro do the [`InnerClass` attribute][1]:
- A class needs one `InnerClass` attribute for each of its nested
classes
- A class needs the `InnerClass` attribute for all (nested) classes
that are mentioned in its constant pool
The attribute for a nested class `A$B$C` consists of the long name of
the outer class `A$B`, the short name of the inner class `C`, and
an access flag set describig the visibility.
The attribute seems to be used for reflection.
[1]: http://docs.oracle.com/javase/specs/jvms/se7/html/jvms-4.html#jvms-4.7.6
|
| | |
|
| | |
|
| |
| |
| |
| |
| | |
No test case for this one, but see the preceding commit for
the class of bug that I'm trying to avoid.
|
|/
|
|
|
| |
It is important to append the fresh 'N' after '$'. Otherwise, we
find out the hard way that ("foo$11" + "1") == ("foo$1" + "11").
|
|\
| |
| | |
Fix checkinit build
|
| | |
|
| |
| |
| |
| | |
Fix only, refactoring in subsequent commit.
|
|\ \ |
|
| |\ \
| | | |
| | | | |
SI-5905 Sanity check -language options
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Underscore means all.
-x:c,b,a,_ results in value c,b,a,a,b,c,d,...
Currently, -Xprint does not present phases as a closed
set of choices; there is ad hoc checking in Global.
That would be a nice unification. (You don't know the
list of choices until after global is constructed.)
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The option names are hardcoded, but checked by a test.
There are no hooks to verify options after the compiler
is constructed.
Introduced a `MultiChoiceSetting` required for the
setting creation framework.
|
|\ \ \ \
| | | | |
| | | | | |
SI-8546 Pattern matcher analysis foiled by over-widening
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
In the enclosed test, the prefix checkable type
`ModuleTypeRef(F2.this, C)` was being inadvertently widened to
`ModuleTypeRef(F2[?], C)`. This started after some misguided
future-proofing in SI-6771 / 3009916.
This commit changes the `dealiasWiden` to a `delias`.
|
|\ \ \ \ \
| | | | | |
| | | | | | |
SI-8531 Better space efficiency for patmat analysis
|
| |/ / / /
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
By adding logging to `clause`, I found that the majority of
calls provide 0 or 1 elements. In SI-7020 / 69557da55, we changed this
method to use a `LinkedHashSet` to have deterministic results
for clauses with more elements. But I suspect that this
contributes to higher memory usage from the pattern matcher.
The enclosed test case, carefully whittled down by @oxbowlakes,
used to consume an inordinate amount of memory and time.
After this patch, it is back to 2.10.4 performance.
I have run `neg/t7020.scala` in a loop and it still is deterministic.
|
|\ \ \ \ \
| | |/ / /
| |/| | | |
|
| |\ \ \ \
| | | | | |
| | | | | | |
makes bundles friendly to -Ywarn-dead-code
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Apparently, the `new Bundle(???).impl` synthetic tree generated as a
macro impl ref for bundles evokes -Ywarn-dead-code warnings.
This pull requests changes `???` to `null` in order not to stress out
the checker. What's in the argument doesn't actually make any difference
anyway.
|
| |\ \ \ \ \
| | | | | | |
| | | | | | | |
SI-8549 Serialization: fix regression with @SerialVersionUID / start enforcing backwards compatibility
|
| | | |/ / /
| | |/| | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
In PR #1673 / 4267444, the annotation `SerialVersionId` was
changed from a `StaticAnnotation` to `ClassFileAnnotation` in
order to avoid silently ignoring non-literal UIDs like:
@SerialVersionUID(0 - 12345L) class C
And to flag non-constant UIDs:
@SerialVersionUID("!!!".length)
While this indeed was fold constants, the change was incomplete.
The compiler API for reading the argument from a `ClassFileAnnoation`
is different, on must look for a `LiteralAnnotArg`, rather than a
`Literal`.
This commit:
- amends the backend accordingly
- removes relevant duplication between `GenASM` and `GenBCode`
- tests that the static field is generated accordingly
This will mean that we will break deserialization of objects from
Scalal 2.11.0 that use this annotation.
|
| |\ \ \ \ \
| | | | | | |
| | | | | | | |
SI-8537 Puts SI-8157 fix under Xsource
|
| | |/ / / / |
|
| |/ / / /
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This is a follow-up to SI-5702 which enabled use of
`*` in infix notation in patterns.
Most of the work is in distinguishing infix from a
sequence pattern.
Also, do not take backticked star as the repeated
parameter marker in postfix position. That is,
`Int``*``` is not `Int*` -- I hope double-tick
renders as tick. There is not a special use case
except that backticks mean "I am an identifier, as
is, and not a keyword."
|
|\ \ \ \ \
| | | | | |
| | | | | | |
SI-7852 for GenBCode
|
| | | | | | |
|
| |/ / / /
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Avoid null checks for "someLiteral".== and SomeModule.==. This has
been implemented in GenICode in #2954.
Introduces a trait to share code between GenICode and GenBCode. This
is just a start, more such refactorings will come quite certainly.
|
|\ \ \ \ \
| | | | | |
| | | | | | |
Merge 2.10.x to 2.11.x
|
| |\ \ \ \ \
| | |/ / / /
| |/| | | | |
|
| | |\ \ \ \
| | | | | | |
| | | | | | | |
SI-8479 Fix constructor default args under scaladoc
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
The `DocDef` node hid the `DefDef` constructor from the scrutinee
of the namer when determining if the class had constructor defaults
or not.
The current pattern for fixing these bugs is to delegate the check
to `TreeInfo`, and account for the wrapper `DocDef` node. I've
followed that pattern, but expressed my feelings about this approach
in a TODO comment.
Before this patch, the enclosed test failed with:
error: not enough arguments for constructor SparkContext: (master: String, appName: String)SparkContext
|